Point-Wise Vibration Pattern Production via a Sparse Actuator Array for Surface Tactile Feedback

Wenbo Ding (丁文伯), Xiaosa Li, Runze Zhao, Chengyue Lu, Xiao Xiao, Wenbo DingICRA 2024 · DOI

TL;DR

This paper presents a novel method for creating precise, localized tactile feedback on a smartphone-sized surface using only five vibration actuators. By modeling actuator interactions and employing a simulated annealing approach to optimize phase and amplitude, they achieve highly accurate point-wise vibration patterns, paving the way for richer haptic experiences in everyday devices.

What Problem Does This Paper Address?

Providing rich, nuanced tactile feedback on handheld electronic devices like smartphones and game controllers is a key goal for enhancing user interaction. While vibration is a common and effective method, densely packing actuators across the entire device surface to create localized feedback is impractical due to size, power, and cost constraints. This limits the ability to convey specific information or create immersive sensations. The challenge, therefore, is to generate precise, point-wise vibration patterns on a surface using a sparse array of actuators, effectively "focusing" the haptic sensation to a specific contact point without requiring a dense actuator grid.

Methods

The researchers developed a tactile feedback board featuring a sparse array of five passive coil actuators. Their approach involves several key steps:

  • Actuator Characterization: They first analyzed the vibration characteristics of a single passive coil actuator with a flip-latch structure. This involved measuring its response to various input waveforms (sine, triangle, square, ramp) and frequencies. They found that a 160 Hz sine wave yielded optimal resonance and energy efficiency.
  • System Modeling: The tactile feedback board was modeled as a Linear Time-Invariant (LTI) system. This model leverages the observed homogeneity and superposition properties of the vibration patterns generated by individual actuators. The displacement of the board is described by an eigenfunction expansion, allowing for the prediction of vibration patterns.
  • Vibration Pattern Synthesis: The core of their method lies in synthesizing composite vibration patterns by linearly combining the individual patterns from the sparse actuator array. This is based on the principle that the vibration patterns from different actuators can be directly summed.
  • Point-Wise Decoupling and Optimization: To achieve precise point-wise feedback, they developed a vibration decoupling process. The goal is to find the optimal phase and amplitude for each of the five actuators such that the resulting composite vibration energy distribution closely matches a target, localized vibration pattern. They employed a simulated annealing algorithm for this optimization. This algorithm iteratively adjusts the actuator parameters, accepting "bad" solutions with a certain probability to avoid getting stuck in local optima, aiming to maximize the Structural Similarity Index Measure (SSIM) between the synthesized and target energy distributions.

Datasets

Not discussed in the paper. The authors generated their own target vibration patterns for evaluation.

Key Insights and Ideas

  • Sparse Actuators, Precise Feedback: The most significant insight is that a small number of strategically placed actuators can indeed generate highly localized tactile feedback, overcoming the need for dense actuator arrays.
  • Vibration as Energy Distribution: The paper highlights that at the optimal frequency (160 Hz), human perception of vibration shifts from waveform variations to sensing an "energy distribution." This insight justifies using the Root Mean Square (RMS) of vibration amplitude as a measure of vibration energy for pattern matching.
  • Simulated Annealing for Haptic Synthesis: The successful application of simulated annealing to find optimal actuator phase spectra for precise haptic feedback is a key methodological contribution. It demonstrates a robust way to solve complex optimization problems in haptic rendering.
  • LTI System Model Simplifies Optimization: Modeling the system as LTI significantly simplifies the problem of synthesizing complex vibration patterns, making the optimization process more tractable.

Key Results

  • High SSIM Scores: The proposed method achieved an average SSIM value of approximately 0.9 when comparing the synthesized point-wise vibration energy patterns to the ideal target patterns. For specific examples, SSIM values ranged from 0.85 to 0.95.
  • Localized Energy Concentration: The synthesized vibration energy patterns showed a clear focal point of energy concentration that closely approximated the target feedback point, with minimal energy spreading to other areas.
  • Efficient Search: The simulated annealing process, while computationally intensive, was able to find optimal solutions within a reasonable timeframe (8-9 hours for a full search), demonstrating the feasibility of real-time or pre-computed feedback generation.

Limitations and Future Work

  • Computational Cost: The simulated annealing algorithm, while effective, can be computationally intensive, requiring significant time for optimization.
  • Pre-computation and Lookup Tables: To address the computational cost for real-time applications, the authors suggest pre-storing optimal phase spectra for various target positions in a lookup table. This would allow for quick retrieval and application based on detected contact points.
  • Generalizability: While the method is demonstrated on a specific tactile board, further research could explore its applicability to different materials, actuator types, and device form factors.
  • Dynamic Interactions: Future work could explore how to dynamically adjust vibration patterns in response to real-time user interactions and changing contact points.

Summary updated

BibTeX
@inproceedings{2024point,
  author = {Wenbo Ding (丁文伯) and Xiaosa Li and Runze Zhao and Chengyue Lu and Xiao Xiao and Wenbo Ding},
  title = {{Point-Wise Vibration Pattern Production via a Sparse Actuator Array for Surface Tactile Feedback}},
  booktitle = {IEEE International Conference on Robotics and Automation},
  year = {2024},
  doi = {10.1109/icra57147.2024.10611362},
}

Loading related content...

Figure